Molecular genetic techniques have been developing rapidly, as have the statistical methods to draw the most information from the new types of results. In human genetics, however, developments on the phenotype side have been less dramatic. The research goal of this proposal is to partially redress that imbalance through the development and application of new methods for the genetic analysis of complex traits in the context of career development for the candidate, Dr. Steven Buyske. The research plan comprises three parts. First, Dr. Buyske will perform secondary analyses of the data collected by the Collaborative Studies on Genetics of Alcoholism, focusing on phenotype definitions involving neuropsychological traits, successful smoking cessation, and successful treatment outcomes for alcohol-dependent subjects, as well as the new methodological approaches developed in this proposal. Second, he will develop a fast, regression-based method for testing genetic linkage of multivariate quantitative traits. Such a method allows several correlated traits to be examined simultaneously. Third, he will develop a novel method for clustering individual diagnostic items to create synthetic phenotypes with a simpler genetic structure. These phenotypes might lack apparent clinical significance yet might bear a closer relationship to the underlying genetic structure, and so would be easier to map. Dr. Buyske will have Robert Pandina, Susan Hodge, David Greenberg, and Linda Brzustowicz as mentors. The training components of this proposal will include coursework in statistical genetics, genetics, alcohol and addiction studies, and the ethical conduct of research, as well as intensive training workshops and consultations. Support through this award will enable Dr. Buyske to develop a unique area of expertise within statistical genetics and develop into an independent investigator capable of conducting important quantitative biomedical research.